Non-functioning pituitary tumours

Non-functioning pituitary tumours

- mortality, morbidity and tumour progression

Daniel S. Olsson

Department of Internal Medicine and Clinical Nutrition Institute of Medicine at the Sahlgrenska Academy

University of Gothenburg Sweden

Gothenburg 2014

Non-functioning pituitary tumours

- mortality, morbidity and tumour progression

Daniel S. Olsson

Department of Internal Medicine and Clinical Nutrition Institute of Medicine at the Sahlgrenska Academy

University of Gothenburg Sweden

Gothenburg 2014

permissions acquired.

Non-functioning pituitary tumours

© Daniel S. Olsson 2014 daniel.olsson@gu.se ISBN 978-91-628-8957-9

E-publication: http://hdl.handle.net/2077/35193 Printed in Gothenburg, Sweden 2014

Printed by Kompendiet/Aidla Trading AB, Gothenburg, Sweden, 2014

permissions acquired.

Non-functioning pituitary tumours

© Daniel S. Olsson 2014 daniel.olsson@gu.se ISBN 978-91-628-8957-9

E-publication: http://hdl.handle.net/2077/35193 Printed in Gothenburg, Sweden 2014

Printed by Kompendiet/Aidla Trading AB, Gothenburg, Sweden, 2014

end of the beginning.”

Sir Winston Churchill (1874-1965), Speech in November 1942

end of the beginning.”

Sir Winston Churchill (1874-1965), Speech in November 1942

- MORTALITY, MORBIDITY AND TUMOUR PROGRESSION Daniel S. Olsson

Department of Internal Medicine and Clinical Nutrition Institute of Medicine at the Sahlgrenska Academy

University of Gothenburg, Sweden

ABSTRACT

Non-functioning pituitary tumours, i.e. non-functioning pituitary adenomas (NFPA) and craniopharyngiomas (CP), are histologically benign brain tumours. They are, however, associated with hypopituitarism, diabetes insipidus and other local symptoms caused by the tumour itself or its treatment. Previous studies have shown an excess mortality in patient populations with hypopituitarism, caused by various aetiologies. The mortality rates and factors predicting the mortality in NFPA and CP patients are largely unknown.

Modern replacement therapy for patients with hypopituitarism includes treatment with growth hormone (GH) replacement therapy (GHRT). GH has known mitogenic effects, and is considered to possibly increase the risk of tumour progression in patients with a history of pituitary tumours.

This thesis is based on four studies aimed to investigate whether GHRT influences the risk of tumour progression and to study mortality and morbidity in patients with NFPA or CP.

In two case-control studies the frequency of tumour progression was investigated in patients with NFPA or CP treated with and without GHRT. The 10-year tumour progression free survival rate in NFPA patients with and without GHRT was 74% and 70%, respectively. The corresponding figures for CP patients were 88% and 57%. In a population-based registry-study of 2795 NFPA patients an excess mortality was demonstrated in women and in patients diagnosed at or before 40 years of age. In another population-based registry-study of 307 CP patients, mortality and morbidity were highly increased, especially in patients with a childhood-onset of the disease. The incidences of type 2 diabetes mellitus, cerebral infarction and severe infection were 5- fold elevated compared to the general population.

In conclusion, GHRT does not affect the frequency of tumour progression in patients with NFPA or CP. Furthermore, there is an increased mortality in women and young patients with NFPA and an excess mortality in CP patients, especially in patients with childhood-onset of CP.

Key words: Non-functioning pituitary adenoma, Craniopharyngioma, Mortality, Morbidity, Growth hormone replacement therapy, Residual tumour, Radiation therapy, Tumour progression

ISBN: 978-91-628-8957-9 (Printed edition) ISBN: 978-91-628-8960-9 (Electronic edition) E-publication: http://hdl.handle.net/2077/35193

- MORTALITY, MORBIDITY AND TUMOUR PROGRESSION Daniel S. Olsson

Department of Internal Medicine and Clinical Nutrition Institute of Medicine at the Sahlgrenska Academy

University of Gothenburg, Sweden

ABSTRACT

Non-functioning pituitary tumours, i.e. non-functioning pituitary adenomas (NFPA) and craniopharyngiomas (CP), are histologically benign brain tumours. They are, however, associated with hypopituitarism, diabetes insipidus and other local symptoms caused by the tumour itself or its treatment. Previous studies have shown an excess mortality in patient populations with hypopituitarism, caused by various aetiologies. The mortality rates and factors predicting the mortality in NFPA and CP patients are largely unknown.

Modern replacement therapy for patients with hypopituitarism includes treatment with growth hormone (GH) replacement therapy (GHRT). GH has known mitogenic effects, and is considered to possibly increase the risk of tumour progression in patients with a history of pituitary tumours.

This thesis is based on four studies aimed to investigate whether GHRT influences the risk of tumour progression and to study mortality and morbidity in patients with NFPA or CP.

In two case-control studies the frequency of tumour progression was investigated in patients with NFPA or CP treated with and without GHRT. The 10-year tumour progression free survival rate in NFPA patients with and without GHRT was 74% and 70%, respectively. The corresponding figures for CP patients were 88% and 57%. In a population-based registry-study of 2795 NFPA patients an excess mortality was demonstrated in women and in patients diagnosed at or before 40 years of age. In another population-based registry-study of 307 CP patients, mortality and morbidity were highly increased, especially in patients with a childhood-onset of the disease. The incidences of type 2 diabetes mellitus, cerebral infarction and severe infection were 5- fold elevated compared to the general population.

In conclusion, GHRT does not affect the frequency of tumour progression in patients with NFPA or CP. Furthermore, there is an increased mortality in women and young patients with NFPA and an excess mortality in CP patients, especially in patients with childhood-onset of CP.

Key words: Non-functioning pituitary adenoma, Craniopharyngioma, Mortality, Morbidity, Growth hormone replacement therapy, Residual tumour, Radiation therapy, Tumour progression

ISBN: 978-91-628-8957-9 (Printed edition)

ISBN: 978-91-628-8960-9 (Electronic edition)

E-publication: http://hdl.handle.net/2077/35193

LIST OF PAPERS

This thesis is based on the following studies, referred to in the text by their Roman numerals.

Paper I. Comparing progression of non-functioning pituitary adenomas in hypopituitarism patients with and without long-term GH replacement therapy.

Olsson DS, Buchfelder M, Schlaffer S, Bengtsson B-Å, Jakobsson K-E, Johannsson G, Nilsson AG.

European Journal of Endocrinology, 2009 161 (5):663-669.

Paper II. Tumour recurrence and enlargement in patients with craniopharyngioma with and without GH replacement therapy during more than 10 years of follow-up.

Olsson DS, Buchfelder M, Wiendieck K, Kremenevskaja N, Bengtsson B-Å, Jakobsson K-E, Jarfelt M, Johannsson G, Nilsson AG.

European Journal of Endocrinology, 2012 166 (6):1061-1068.

Paper III. Mortality in patients with non-functioning pituitary adenoma – a population-based study.

Olsson DS, Johannsson G, Bryngelsson I-L, Trimpou P, Nilsson AG, Andersson E.

Manuscript.

Paper IV. Mortality and morbidity in patients with craniopharyngioma – a population-based study.

Olsson DS, Andersson E, Bryngelsson I-L, Nilsson AG, Johannsson G.

Manuscript.

All papers have been accepted for oral presentation at The Endocrine Society´s Annual Meeting (Paper I, 2009, Washington, USA; Paper II, 2011, Boston, USA; Paper III and Paper IV, 2014, Chicago, USA).

LIST OF PAPERS

This thesis is based on the following studies, referred to in the text by their Roman numerals.

Paper I. Comparing progression of non-functioning pituitary adenomas in hypopituitarism patients with and without long-term GH replacement therapy.

Olsson DS, Buchfelder M, Schlaffer S, Bengtsson B-Å, Jakobsson K-E, Johannsson G, Nilsson AG.

European Journal of Endocrinology, 2009 161 (5):663-669.

Paper II. Tumour recurrence and enlargement in patients with craniopharyngioma with and without GH replacement therapy during more than 10 years of follow-up.

Olsson DS, Buchfelder M, Wiendieck K, Kremenevskaja N, Bengtsson B-Å, Jakobsson K-E, Jarfelt M, Johannsson G, Nilsson AG.

European Journal of Endocrinology, 2012 166 (6):1061-1068.

Paper III. Mortality in patients with non-functioning pituitary adenoma – a population-based study.

Olsson DS, Johannsson G, Bryngelsson I-L, Trimpou P, Nilsson AG, Andersson E.

Manuscript.

Paper IV. Mortality and morbidity in patients with craniopharyngioma – a population-based study.

Olsson DS, Andersson E, Bryngelsson I-L, Nilsson AG, Johannsson G.

Manuscript.

All papers have been accepted for oral presentation at The Endocrine Society´s Annual

Meeting (Paper I, 2009, Washington, USA; Paper II, 2011, Boston, USA; Paper III

and Paper IV, 2014, Chicago, USA).

TABLE OF CONTENTS

A BBREVIATIONS ... 4

1 G ENERAL I NTRODUCTION ... 5

1.1 History ... 5

1.2 Pathogenesis and epidemiology ... 5

1.3 Symptoms and signs ... 6

1.4 Tumour treatment ... 6

1.5 Hypopituitarism and growth hormone deficiency ... 7

1.6 Mitogenic effects of GH and IGF-I ... 7

1.7 Tumour progression ... 9

1.7.1 Tumour progression in patients with NFPA ... 9

1.7.2 Tumour progression in patients with CP ... 10

1.8 Mortality and morbidity ... 12

1.8.1 Mortality in patients with NFPA ... 12

1.8.2 Mortality and morbidity in patients with CP ... 13

2 A IM ... 14

3 S UBJECTS AND M ETHODS ... 15

3.1 Study design and subjects ... 15

3.1.1 Paper I ... 15

3.1.2 Paper II ... 16

3.1.3 Paper III ... 17

3.1.4 Paper IV ... 18

3.2 Statistical methods ... 20

3.2.1 Paper I-II ... 20

3.2.2 Paper III-IV ... 20

4 R ESULTS ... 21

4.1 Paper I ... 21

4.1.1 Patients treated with GHRT ... 21

4.1.2 Control patients ... 21

4.1.3 Comparison between patients with and without GHRT ... 22

4.2 Paper II ... 22

TABLE OF CONTENTS A BBREVIATIONS ... 4

1 G ENERAL I NTRODUCTION ... 5

1.1 History ... 5

1.2 Pathogenesis and epidemiology ... 5

1.3 Symptoms and signs ... 6

1.4 Tumour treatment ... 6

1.5 Hypopituitarism and growth hormone deficiency ... 7

1.6 Mitogenic effects of GH and IGF-I ... 7

1.7 Tumour progression ... 9

1.7.1 Tumour progression in patients with NFPA ... 9

1.7.2 Tumour progression in patients with CP ... 10

1.8 Mortality and morbidity ... 12

1.8.1 Mortality in patients with NFPA ... 12

1.8.2 Mortality and morbidity in patients with CP ... 13

2 A IM ... 14

3 S UBJECTS AND M ETHODS ... 15

3.1 Study design and subjects ... 15

3.1.1 Paper I ... 15

3.1.2 Paper II ... 16

3.1.3 Paper III ... 17

3.1.4 Paper IV ... 18

3.2 Statistical methods ... 20

3.2.1 Paper I-II ... 20

3.2.2 Paper III-IV ... 20

4 R ESULTS ... 21

4.1 Paper I ... 21

4.1.1 Patients treated with GHRT ... 21

4.1.2 Control patients ... 21

4.1.3 Comparison between patients with and without GHRT ... 22

4.2 Paper II ... 22

4.2.1 Long-term outcome for all patients ... 22

4.2.2 Comparison between patients with and without GHRT... 23

4.3 Paper III... 24

4.3.1 Overall mortality ... 24

4.3.2 Cause specific mortality... 26

4.4 Paper IV... 26

4.4.1 Mortality ... 26

4.4.2 Morbidity ... 26

5 G ENERAL D ISCUSSION ... 28

5.1 Epidemiology ... 28

5.2 Frequency of hormonal deficiencies... 28

5.3 Tumour progression... 29

5.3.1 Tumour progression and GHRT... 29

5.4 Mortality and morbidity ... 30

5.5 Methodological considerations... 32

5.5.1 Paper I-II ... 32

5.5.2 Paper III-IV... 33

5.5.3 Overlapping study populations ... 34

6 C ONCLUSION ... 35

7 F UTURE PERSPECTIVES ... 36

S AMMANFATTNING PÅ SVENSKA ... 37

A CKNOWLEDGEMENTS ... 38

R EFERENCES ... 40

P APERS ... 49

4.2.1 Long-term outcome for all patients ... 22

4.2.2 Comparison between patients with and without GHRT... 23

4.3 Paper III... 24

4.3.1 Overall mortality ... 24

4.3.2 Cause specific mortality... 26

4.4 Paper IV... 26

4.4.1 Mortality ... 26

4.4.2 Morbidity ... 26

5 G ENERAL D ISCUSSION ... 28

5.1 Epidemiology ... 28

5.2 Frequency of hormonal deficiencies... 28

5.3 Tumour progression... 29

5.3.1 Tumour progression and GHRT... 29

5.4 Mortality and morbidity ... 30

5.5 Methodological considerations... 32

5.5.1 Paper I-II ... 32

5.5.2 Paper III-IV... 33

5.5.3 Overlapping study populations ... 34

6 C ONCLUSION ... 35

7 F UTURE PERSPECTIVES ... 36

S AMMANFATTNING PÅ SVENSKA ... 37

A CKNOWLEDGEMENTS ... 38

R EFERENCES ... 40

P APERS ... 49

ABBREVIATIONS

Cancer Registry The Swedish Cancer Registry

Cause of Death Registry The Swedish National Cause of Death Registry

CI Confidence interval

CP Craniopharyngioma

DI Diabetes insipidus

GH Growth hormone

GHD Growth hormone deficiency

GHRT Growth hormone replacement therapy

HR Hazard ratio

IGF-I Insulin-like growth factor - I

IGFBP-3 Insulin-like growth factor-binding protein - 3

MRI Magnetic resonance imaging

NFPA Non-functioning pituitary adenoma

RT Radiation therapy

SD Standard deviation

SIR Standardised incidence ratio

SMR Standardised mortality ratio

PFSR Progression free survival rate

Patient Registry The Swedish National Patient Registry

T2DM Type 2 diabetes mellitus

ABBREVIATIONS

Cancer Registry The Swedish Cancer Registry

Cause of Death Registry The Swedish National Cause of Death Registry

CI Confidence interval

CP Craniopharyngioma

DI Diabetes insipidus

GH Growth hormone

GHD Growth hormone deficiency

GHRT Growth hormone replacement therapy

HR Hazard ratio

IGF-I Insulin-like growth factor - I

IGFBP-3 Insulin-like growth factor-binding protein - 3

MRI Magnetic resonance imaging

NFPA Non-functioning pituitary adenoma

RT Radiation therapy

SD Standard deviation

SIR Standardised incidence ratio

SMR Standardised mortality ratio

PFSR Progression free survival rate

Patient Registry The Swedish National Patient Registry

T2DM Type 2 diabetes mellitus

1 GENERAL INTRODUCTION

1.1 History

It is unclear when the first pituitary tumour was described, but the Austrian physician A. de Haen (1704-1776) is believed to be the first to link a pituitary mass to symptoms in a female patient with amenorrhoea [1]. More than one hundred years later, O. Minkowski proposed that abnormalities in the pituitary gland could be associated with a disease - acromegaly [2]. A few years later, in 1909, H. Cushing suggested that tumours in the pituitary gland could lead to hypo- or hyper-secretion of hormones [3].

One of the first to describe the entity of what later would be known as craniopharyngioma (CP) was J. Erdheim, when he in 1904 presented a series of ten patients with nests of squamous epithelium at the junction between the infundibulum and the pituitary [4]. Almost thirty years later, in 1932, H. Cushing was the first to use the designation craniopharyngioma [5]. Since then, the knowledge about the pituitary and its associated tumours has grown enormously and has today expanded into an independent area of research.

1.2 Pathogenesis and epidemiology

Today we know that pituitary adenomas originate from cells in the pituitary gland and consist of secreting pituitary adenomas and non-functioning pituitary adenomas (NFPA). Pituitary adenomas are a tumour group with different aetiologies consisting of prolactinomas (26-51%), NFPAs (27-36%), GH-producing adenomas (9-16%), ACTH- producing adenomas (3-15%) and TSHomas (1%) [6-8]. CP is an epithelial tumour, whose origin is uncertain [9]. Several theories on the aetiology of CP have been suggested, but one hypothesis is that CP tumours arise from metaplasia of adenohypophyseal cells in the pituitary stalk or gland [10,11]. CP tumours have been demonstrated to have two different primary histological appearances, the adamantinomatous and the papillary subtype, but there are also cases with a mixture of these subtypes [12]. NFPAs and CPs have only rarely been described to transform into malignant tumours [13,14].

The annual incidence of NFPA has been reported to be 1.0-1.1 cases per 100 000 inhabitants in two studies from the Nordic countries [6,15]. Mean age at diagnosis is 50- 54 years [15-17], with no difference between men and women [16]. CP, which is a less common tumour, has an annual incidence of 0.13-0.17 cases per 100 000 inhabitants [18,19] but constitute 5-15% of intracranial tumours in children [20,21]. In patients with CP, the age at diagnosis has been reported to have a bimodal shape, with one peak during childhood and another during adulthood [18,22,23].

1 GENERAL INTRODUCTION

1.1 History

It is unclear when the first pituitary tumour was described, but the Austrian physician A. de Haen (1704-1776) is believed to be the first to link a pituitary mass to symptoms in a female patient with amenorrhoea [1]. More than one hundred years later, O. Minkowski proposed that abnormalities in the pituitary gland could be associated with a disease - acromegaly [2]. A few years later, in 1909, H. Cushing suggested that tumours in the pituitary gland could lead to hypo- or hyper-secretion of hormones [3].

One of the first to describe the entity of what later would be known as craniopharyngioma (CP) was J. Erdheim, when he in 1904 presented a series of ten patients with nests of squamous epithelium at the junction between the infundibulum and the pituitary [4]. Almost thirty years later, in 1932, H. Cushing was the first to use the designation craniopharyngioma [5]. Since then, the knowledge about the pituitary and its associated tumours has grown enormously and has today expanded into an independent area of research.

1.2 Pathogenesis and epidemiology

Today we know that pituitary adenomas originate from cells in the pituitary gland and consist of secreting pituitary adenomas and non-functioning pituitary adenomas (NFPA). Pituitary adenomas are a tumour group with different aetiologies consisting of prolactinomas (26-51%), NFPAs (27-36%), GH-producing adenomas (9-16%), ACTH- producing adenomas (3-15%) and TSHomas (1%) [6-8]. CP is an epithelial tumour, whose origin is uncertain [9]. Several theories on the aetiology of CP have been suggested, but one hypothesis is that CP tumours arise from metaplasia of adenohypophyseal cells in the pituitary stalk or gland [10,11]. CP tumours have been demonstrated to have two different primary histological appearances, the adamantinomatous and the papillary subtype, but there are also cases with a mixture of these subtypes [12]. NFPAs and CPs have only rarely been described to transform into malignant tumours [13,14].

The annual incidence of NFPA has been reported to be 1.0-1.1 cases per 100 000

inhabitants in two studies from the Nordic countries [6,15]. Mean age at diagnosis is 50-

54 years [15-17], with no difference between men and women [16]. CP, which is a less

common tumour, has an annual incidence of 0.13-0.17 cases per 100 000 inhabitants

[18,19] but constitute 5-15% of intracranial tumours in children [20,21]. In patients with

CP, the age at diagnosis has been reported to have a bimodal shape, with one peak

during childhood and another during adulthood [18,22,23].

1.3 Symptoms and signs

The manifestations of a non-functioning pituitary tumour are determined by tumour size, the velocity of tumour growth and the direction of the tumour expansion. The presenting symptoms are often visual field defects or symptoms related to hormonal deficiencies but can also be headache, symptoms from pressure on cranial nerves, rhinorrhoea or nausea [16,17,24-29]. In some cases the presenting symptom can be a haemorrhage or an infarction of the adenoma, i.e. tumour apoplexy, which can lead to sudden expansion of the tumour resulting in acute headache, visual impairment and hypopituitarism [30]. It is not unusual that a pituitary mass causes a bitemporal hemianopsia due to pressure on the central optic chiasm of the tumour mass. Depending on the duration of the pressure an optic atrophy can develop leading to irreversible visual field defects and reduced acuity. For patients with childhood-onset of a pituitary tumour (mainly CP) failure of growth due to hypopituitarism can be an early symptom leading to further medical evaluation [27-29]. Early symptoms in women with pituitary masses are amenorrhoea, anovulation and infertility [16,25] that can lead the investigation into examining the function of the pituitary gland.

1.4 Tumour treatment

After the introduction of magnetic resonance imaging (MRI) the number of incidentally found pituitary tumours, especially NFPA, has increased. This has increased the number of patients with NFPA that can be followed with a wait and see strategy. The natural history of NFPAs is to some degree unclear. The existing studies have been small and investigated incidentally found tumours or tumours leading to symptoms for which surgery was postponed for various reasons [31-34].

The therapeutic options for patients with a non-functioning pituitary tumour are surgery, radiation therapy (RT), medical therapy and conservative treatment. Individual factors, such as severity of symptoms (especially optic nerve compression) as well as tumour type and location, influence the choice of treatment. If a pituitary tumour threatens the optic chiasm, the primary choice of tumour treatment is surgery [31,33]. When surgical treatment is performed, the first-hand choice is a transsphenoidal approach, whereas the transcranial approach is used in complicated cases and when the tumour cannot be resected from inside the sella. In addition to the nature of the tumour, the outcome of surgery is highly dependent on the experience of the surgeon [35].

For CP tumours the primary treatment is surgery. In CP tumours that are predominantly cystic the resection can be facilitated by preoperative fluid aspiration. The question of whether gross total remove should be attempted has been debated. Today many centres try to perform a gross total removal as long as hazardous manipulations of critical brain areas can be avoided [36]. Several factors, including firm attachment of the tumour capsule to blood vessels or hypothalamic tissue, may inhibit the surgeon from executing a total gross removal and instead aim for a partial removal and preservation of the vessels and the hypothalamus [36,37].

1.3 Symptoms and signs

The manifestations of a non-functioning pituitary tumour are determined by tumour size, the velocity of tumour growth and the direction of the tumour expansion. The presenting symptoms are often visual field defects or symptoms related to hormonal deficiencies but can also be headache, symptoms from pressure on cranial nerves, rhinorrhoea or nausea [16,17,24-29]. In some cases the presenting symptom can be a haemorrhage or an infarction of the adenoma, i.e. tumour apoplexy, which can lead to sudden expansion of the tumour resulting in acute headache, visual impairment and hypopituitarism [30]. It is not unusual that a pituitary mass causes a bitemporal hemianopsia due to pressure on the central optic chiasm of the tumour mass. Depending on the duration of the pressure an optic atrophy can develop leading to irreversible visual field defects and reduced acuity. For patients with childhood-onset of a pituitary tumour (mainly CP) failure of growth due to hypopituitarism can be an early symptom leading to further medical evaluation [27-29]. Early symptoms in women with pituitary masses are amenorrhoea, anovulation and infertility [16,25] that can lead the investigation into examining the function of the pituitary gland.

1.4 Tumour treatment

After the introduction of magnetic resonance imaging (MRI) the number of incidentally found pituitary tumours, especially NFPA, has increased. This has increased the number of patients with NFPA that can be followed with a wait and see strategy. The natural history of NFPAs is to some degree unclear. The existing studies have been small and investigated incidentally found tumours or tumours leading to symptoms for which surgery was postponed for various reasons [31-34].

The therapeutic options for patients with a non-functioning pituitary tumour are surgery, radiation therapy (RT), medical therapy and conservative treatment. Individual factors, such as severity of symptoms (especially optic nerve compression) as well as tumour type and location, influence the choice of treatment. If a pituitary tumour threatens the optic chiasm, the primary choice of tumour treatment is surgery [31,33]. When surgical treatment is performed, the first-hand choice is a transsphenoidal approach, whereas the transcranial approach is used in complicated cases and when the tumour cannot be resected from inside the sella. In addition to the nature of the tumour, the outcome of surgery is highly dependent on the experience of the surgeon [35].

For CP tumours the primary treatment is surgery. In CP tumours that are predominantly

cystic the resection can be facilitated by preoperative fluid aspiration. The question of

whether gross total remove should be attempted has been debated. Today many centres

try to perform a gross total removal as long as hazardous manipulations of critical brain

areas can be avoided [36]. Several factors, including firm attachment of the tumour

capsule to blood vessels or hypothalamic tissue, may inhibit the surgeon from executing

a total gross removal and instead aim for a partial removal and preservation of the

vessels and the hypothalamus [36,37].

In most cases when total tumour removal is not possible or when is not attempted, the second line of treatment is RT. RT treatment is also used in recurrent NFPA and CP tumours when surgical treatment is no longer an option. Although treatment with RT has shown a high rate of tumour control [38-41], it is also associated with an increased risk of developing additional pituitary hormone deficiencies, an increased risk of optic neuropathy as well as secondary brain tumours [42-46]. In addition, RT for pituitary tumours has been associated with an increased morbidity and mortality due to cerebrovascular diseases [47-49]. The combination of these side effects and a suggested increased cerebrovascular mortality has resulted in some centres using RT more conservatively in the general treatment of NFPA and CP patients during the last decade.

1.5 Hypopituitarism and growth hormone deficiency

Hypopituitarism is defined as a non-adequate secretion of one or several of the hormones produced by the anterior lobe of the pituitary gland. Many patients with a non-functioning pituitary tumour develop hypopituitarism, due to the volume effect of the tumour or as a consequence of tumour treatment. Somatotropin deficiency, i.e.

growth hormone (GH) deficiency (GHD), is often the first hormone deficiency to occur and is present in almost all patients with a pituitary tumour who have received RT and in 60-80% of patients treated with surgery alone as tumour treatment [42,50,51]. These patients also often develop additional hormone deficiencies, i.e. gonadotropin, thyrotropin and/or cortotropin deficiency. Diabetes insipidus (DI) is less common (0- 10%) in patients with NFPA [16,17,52,53] but frequent in patients with CP, especially after surgical treatment when approximately 60% of CP patients suffer from DI [51,54].

Historically all hormone deficiencies excluding GHD have been treated with replacement therapy. The clinical picture of non-replaced GHD in hypopituitary adults was recognised in the beginning of the 1990s and it has been associated with increased vascular mortality, premature atherosclerosis, abnormal body composition, reduced bone mass, an unfavourable lipid profile, reduced muscle strength and fatigue and reduced quality of life [55-62]. The paediatric use of GH therapy started already in the 1950s, when physicians first treated children with short stature and severe GHD [63].

Since then, many children have been able to reach normal or near normal final height with the help of GH therapy. In addition to enabling short statured children to reach full height, GH replacement therapy (GHRT) has been shown to improve most of the abnormalities in adult GHD [64-67]. Long-term GHRT has therefore become common practice in many countries. The efficacy and safety profile of GHRT has led to a continuation of long-term GHRT in adults and continued therapy of GHD children into adulthood if GHD is still present.

1.6 Mitogenic effects of GH and IGF-I

Already in the 1950s, it was suggested that GH had a role in the development of malignant tumours, since hypophysectomy was seen to induce remission in patients with metastatic mammary carcinoma [68]. Today, after more than half a century, circulating and extra pituitary expression of GH and insulin-like growth factor-I (IGF-I) has been proposed to have a roll in the development and progression of tumours. The

In most cases when total tumour removal is not possible or when is not attempted, the second line of treatment is RT. RT treatment is also used in recurrent NFPA and CP tumours when surgical treatment is no longer an option. Although treatment with RT has shown a high rate of tumour control [38-41], it is also associated with an increased risk of developing additional pituitary hormone deficiencies, an increased risk of optic neuropathy as well as secondary brain tumours [42-46]. In addition, RT for pituitary tumours has been associated with an increased morbidity and mortality due to cerebrovascular diseases [47-49]. The combination of these side effects and a suggested increased cerebrovascular mortality has resulted in some centres using RT more conservatively in the general treatment of NFPA and CP patients during the last decade.

1.5 Hypopituitarism and growth hormone deficiency

Hypopituitarism is defined as a non-adequate secretion of one or several of the hormones produced by the anterior lobe of the pituitary gland. Many patients with a non-functioning pituitary tumour develop hypopituitarism, due to the volume effect of the tumour or as a consequence of tumour treatment. Somatotropin deficiency, i.e.

growth hormone (GH) deficiency (GHD), is often the first hormone deficiency to occur and is present in almost all patients with a pituitary tumour who have received RT and in 60-80% of patients treated with surgery alone as tumour treatment [42,50,51]. These patients also often develop additional hormone deficiencies, i.e. gonadotropin, thyrotropin and/or cortotropin deficiency. Diabetes insipidus (DI) is less common (0- 10%) in patients with NFPA [16,17,52,53] but frequent in patients with CP, especially after surgical treatment when approximately 60% of CP patients suffer from DI [51,54].

Historically all hormone deficiencies excluding GHD have been treated with replacement therapy. The clinical picture of non-replaced GHD in hypopituitary adults was recognised in the beginning of the 1990s and it has been associated with increased vascular mortality, premature atherosclerosis, abnormal body composition, reduced bone mass, an unfavourable lipid profile, reduced muscle strength and fatigue and reduced quality of life [55-62]. The paediatric use of GH therapy started already in the 1950s, when physicians first treated children with short stature and severe GHD [63].

Since then, many children have been able to reach normal or near normal final height with the help of GH therapy. In addition to enabling short statured children to reach full height, GH replacement therapy (GHRT) has been shown to improve most of the abnormalities in adult GHD [64-67]. Long-term GHRT has therefore become common practice in many countries. The efficacy and safety profile of GHRT has led to a continuation of long-term GHRT in adults and continued therapy of GHD children into adulthood if GHD is still present.

1.6 Mitogenic effects of GH and IGF-I

Already in the 1950s, it was suggested that GH had a role in the development of

malignant tumours, since hypophysectomy was seen to induce remission in patients

with metastatic mammary carcinoma [68]. Today, after more than half a century,

circulating and extra pituitary expression of GH and insulin-like growth factor-I (IGF-I)

has been proposed to have a roll in the development and progression of tumours. The

evidence is based on animal models where different types of transgenic mice with reduced levels of GH and IGF-I are resistant to carcinogenesis induced by chemicals [69-71]. In contrast, transgenic mice with excess human-GH or IGF-I have shown an increased rate of mammary tumours or epidermal tumours [72-74]. Since GH and IGF-I are mitogenic factors and studies have found GH receptors in both NFPA and CP cells [75-77], there is a theoretical risk of an increased frequency of tumour progression in patients treated with long-term GHRT [78,79].

Due to the clinical benefits in patients with GHD, long-term GHRT has become common practice in many countries. There are, for reason mentioned above, some concerns regarding the possibility that GH and IGF-I may be associated with cancer and an elevated risk of tumour progression. Patients with acromegaly who, often for many years, have supraphysiological levels of GH and IGF-I have been studied with conflicting results. Some studies have shown an increased risk of cancer in acromegalic patients [80-83] whereas other studies have not been able to demonstrate this [84,85].

These conflicting data cannot rule out that the elevated GH and IGF-I concentrations may play a role in the risk of cancer. Popovic et al. showed an increased incidence of neoplasia in acromegalic patients, although the incidence was not higher than in patients with NFPA [86]. Another study described an increased incidence of colorectal cancer in patients with hypopituitarism without GHRT [87]. These results suggest that there may be another explanation for the increased incidence of malignancies in acromegaly, such as an inherently higher risk in hypopituitarism, or that patients with pituitary adenomas are prone to develop neoplasia in general, or the use of RT in this patient population [45,87]. It is important to remember when comparing GHRT to acromegaly that the aim of GHRT is to obtain physiological levels of GH and not supra-physiological levels as in acromegaly.

Several studies have investigated the association between serum IGF-I concentrations and malignancies in the normal population. For instance, Chan et al. have shown a positive association between prostate cancer risk and serum IGF-I concentrations in the highest quartile of the normal range [88]. In a meta-analysis of the association between concentrations of IGF-I and insulin-like growth factor-binding protein-3 (IGFBP-3) with prostate cancer, colorectal cancer, breast cancer and lung cancer, high circulating concentrations of IGF-I or IGFBP-3 were associated with an increased risk of prostate cancer, colorectal cancer and premenopausal breast cancer [89]. These associations were modest and vary between tumour sites. In addition, the frequency of secondary neoplasms was not elevated in a large population of adult GHD patients treated with long-term GHRT [90].

Since GH and IGF-I are known mitogenic factors, there is a theoretical risk that GHRT could potentially increase the rate of tumour progression in patients with pituitary tumours. This is an important safety issue for the majority of hypopituitary patients who have an underlying pituitary tumour. Occasionally case reports appear regarding patients with rapid growth of a pituitary tumour after initiation of GHRT [91,92]. In addition it has been shown that patients who develop progression of a pituitary tumour have an excess mortality [93]. Thus, the question of whether GHRT increases the

evidence is based on animal models where different types of transgenic mice with reduced levels of GH and IGF-I are resistant to carcinogenesis induced by chemicals [69-71]. In contrast, transgenic mice with excess human-GH or IGF-I have shown an increased rate of mammary tumours or epidermal tumours [72-74]. Since GH and IGF-I are mitogenic factors and studies have found GH receptors in both NFPA and CP cells [75-77], there is a theoretical risk of an increased frequency of tumour progression in patients treated with long-term GHRT [78,79].

Due to the clinical benefits in patients with GHD, long-term GHRT has become common practice in many countries. There are, for reason mentioned above, some concerns regarding the possibility that GH and IGF-I may be associated with cancer and an elevated risk of tumour progression. Patients with acromegaly who, often for many years, have supraphysiological levels of GH and IGF-I have been studied with conflicting results. Some studies have shown an increased risk of cancer in acromegalic patients [80-83] whereas other studies have not been able to demonstrate this [84,85].

These conflicting data cannot rule out that the elevated GH and IGF-I concentrations may play a role in the risk of cancer. Popovic et al. showed an increased incidence of neoplasia in acromegalic patients, although the incidence was not higher than in patients with NFPA [86]. Another study described an increased incidence of colorectal cancer in patients with hypopituitarism without GHRT [87]. These results suggest that there may be another explanation for the increased incidence of malignancies in acromegaly, such as an inherently higher risk in hypopituitarism, or that patients with pituitary adenomas are prone to develop neoplasia in general, or the use of RT in this patient population [45,87]. It is important to remember when comparing GHRT to acromegaly that the aim of GHRT is to obtain physiological levels of GH and not supra-physiological levels as in acromegaly.

Several studies have investigated the association between serum IGF-I concentrations and malignancies in the normal population. For instance, Chan et al. have shown a positive association between prostate cancer risk and serum IGF-I concentrations in the highest quartile of the normal range [88]. In a meta-analysis of the association between concentrations of IGF-I and insulin-like growth factor-binding protein-3 (IGFBP-3) with prostate cancer, colorectal cancer, breast cancer and lung cancer, high circulating concentrations of IGF-I or IGFBP-3 were associated with an increased risk of prostate cancer, colorectal cancer and premenopausal breast cancer [89]. These associations were modest and vary between tumour sites. In addition, the frequency of secondary neoplasms was not elevated in a large population of adult GHD patients treated with long-term GHRT [90].

Since GH and IGF-I are known mitogenic factors, there is a theoretical risk that GHRT

could potentially increase the rate of tumour progression in patients with pituitary

tumours. This is an important safety issue for the majority of hypopituitary patients who

have an underlying pituitary tumour. Occasionally case reports appear regarding

patients with rapid growth of a pituitary tumour after initiation of GHRT [91,92]. In

addition it has been shown that patients who develop progression of a pituitary tumour

have an excess mortality [93]. Thus, the question of whether GHRT increases the

tumour progression rate is important not only for achieving local tumour control but also for the overall outcome in these patients.

1.7 Tumour progression

Tumour progression in non-functioning pituitary tumours is often reported as tumour progression free survival rate (PFSR). Most studies define tumour progression as any increase in the tumour size depicted at imaging regardless of clinical symptoms or need of additional tumour treatment. The PFSR varies between NFPAs and CPs and is therefore described separately.

1.7.1 Tumour progression in patients with NFPA

Tumour progression rate may be underestimated in older studies when computerised tomography was used instead of MRI, which is routinely used in the management of pituitary tumours today. The 10-year PFSR in NFPA patients whose tumour treatment involved RT has been reported to be 89%-98% [25,38,39,94] (Table 1). In patients treated with surgery alone, the 10-year PFSR for patients with and without a residual tumour after the primary tumour treatment was 23%-58% and 91%-100%, respectively [25,95]. Thus, the risk of tumour progression is strongly dependent on the outcome of the primary surgical treatment as well as on whether RT was part of the primary tumour treatment [16,24,25].

Three early studies on patients with pituitary tumours, including those with NFPA, who received GHRT, did not report an increase in the tumour progression rate. They indeed had low progression rates compared to historical series [58,96,97] (Table 1). These studies had severe limitations as the tumours had mixed aetiologies, there was also a short follow-up, no comparison with a control group and, in some a high percentage of initial RT. These factors may confound detection of any negative effect of GHRT, in particular on a slowly growing tumour such as NFPA.

In the first comparative study of 55 NFPA patients with GHRT, Buchfelder et al.

showed a similar frequency of tumour progression in patients with and without GHRT [98] (Table 1). The optimal setting for investigating any stimulating effect of GHRT on tumour cells is to study RT naive NFPAs, since RT has a well-known anti-proliferative effect. Arnold and colleagues studied the effect of GHRT on 23 RT naive NFPA patients receiving GHRT for a mean duration of 4.6 years and did not find that GHRT was an independent predictor of tumour progression [99]. In summary, larger comparative studies of NFPA patients with longer follow-up are still needed to unravel whether GHRT is associated with an increased risk of tumour progression.

tumour progression rate is important not only for achieving local tumour control but also for the overall outcome in these patients.

1.7 Tumour progression

Tumour progression in non-functioning pituitary tumours is often reported as tumour progression free survival rate (PFSR). Most studies define tumour progression as any increase in the tumour size depicted at imaging regardless of clinical symptoms or need of additional tumour treatment. The PFSR varies between NFPAs and CPs and is therefore described separately.

1.7.1 Tumour progression in patients with NFPA

Tumour progression rate may be underestimated in older studies when computerised tomography was used instead of MRI, which is routinely used in the management of pituitary tumours today. The 10-year PFSR in NFPA patients whose tumour treatment involved RT has been reported to be 89%-98% [25,38,39,94] (Table 1). In patients treated with surgery alone, the 10-year PFSR for patients with and without a residual tumour after the primary tumour treatment was 23%-58% and 91%-100%, respectively [25,95]. Thus, the risk of tumour progression is strongly dependent on the outcome of the primary surgical treatment as well as on whether RT was part of the primary tumour treatment [16,24,25].

Three early studies on patients with pituitary tumours, including those with NFPA, who received GHRT, did not report an increase in the tumour progression rate. They indeed had low progression rates compared to historical series [58,96,97] (Table 1). These studies had severe limitations as the tumours had mixed aetiologies, there was also a short follow-up, no comparison with a control group and, in some a high percentage of initial RT. These factors may confound detection of any negative effect of GHRT, in particular on a slowly growing tumour such as NFPA.

In the first comparative study of 55 NFPA patients with GHRT, Buchfelder et al.

showed a similar frequency of tumour progression in patients with and without GHRT

[98] (Table 1). The optimal setting for investigating any stimulating effect of GHRT on

tumour cells is to study RT naive NFPAs, since RT has a well-known anti-proliferative

effect. Arnold and colleagues studied the effect of GHRT on 23 RT naive NFPA

patients receiving GHRT for a mean duration of 4.6 years and did not find that GHRT

was an independent predictor of tumour progression [99]. In summary, larger

comparative studies of NFPA patients with longer follow-up are still needed to unravel

whether GHRT is associated with an increased risk of tumour progression.

Table 1. Tumour progression free survival in NFPA patients sub-grouped depending on GHRT.

Study – First author [ref]

No. of patients

Aetiology Mean follow-up (yrs)

RT Progression free survival rate:

10 years

All patients treated with GHRT §

Frajese [96] 100 Mixed 2.8^ 91% (NED) 1 patient had

tumour progression

Abs [58] 1034 Mixed 0.8 No

data

(NED) 6 patients had tumour progression

Chung [97] 48 Mixed 3.2¤ 0% (NED) 3 patients had

tumour progression Comparison between patients with and without GHRT §

Hatrick [100] 47 (GHRT) Mixed 3.6 38% (NED) 2 patients had tumour progression 28 (No GH) Mixed 3.4 75% (NED) 2 patients had

tumour progression Buchfelder [98] 55 (GHRT) All NFPA 5 (At least) 20% (NED) 29% had

tumour progression 55 (No GH) All NFPA 5 (At least) 24% (NED) 22% had

tumour progression

Arnold [99] 23 (GHRT) All NFPA 4.6 0% (NED) 35% had

tumour progression 107 (No GH) All NFPA 6.8 0% (NED) 36% had

tumour progression Reference series in patients without GHRT

Gitteos [38] 63 All NFPA 8.1 0% 47%

63 All NFPA 9.1 100% 93%

Brada [94] 252 All NFPA 10.5¤ 100% 94%

Flickinger [101] 112 All NFPA No data 100% 89%

Park [39] 132 All NFPA 3.8¤ 0% 49%

44 All NFPA 5.7¤ 100% 98%

O’Sullivan [95] 126 All NFPA 5.7¤ 0% 49%

GHRT, Growth hormone replacement therapy, NED, No exact data; NFPA, Non-functioning pituitary adenoma;

RT, Radiation therapy; ^, Calculated number; ¤, Follow-up period presented as median value; §, For patients treated with GHRT the mean follow-up time refers to the mean duration of GHRT.

1.7.2 Tumour progression in patients with CP

Craniopharyngiomas have a strong tendency to progress and infiltrate the surrounding structures, which affects the choice and aggressiveness of tumour treatment. As for NFPAs, the progression rate for CP is strongly affected by the inclusion of RT in the primary treatment. The 10-year PFSR in CP patients with initial RT treatment has been reported to be 83%-92% [27,41,102] (Table 2). In CP patients treated with surgery alone the 10-year PFSR was 41% for patients with a residual tumour after primary tumour treatment [27]. The corresponding figure for patients treated surgically without any residual tumour was 47%-81% [27,36,102].

Table 1. Tumour progression free survival in NFPA patients sub-grouped depending on GHRT.

Study – First author [ref]

No. of patients

Aetiology Mean follow-up (yrs)

RT Progression free survival rate:

10 years

All patients treated with GHRT §

Frajese [96] 100 Mixed 2.8^ 91% (NED) 1 patient had

tumour progression

Abs [58] 1034 Mixed 0.8 No

data

(NED) 6 patients had tumour progression

Chung [97] 48 Mixed 3.2¤ 0% (NED) 3 patients had

tumour progression Comparison between patients with and without GHRT §

Hatrick [100] 47 (GHRT) Mixed 3.6 38% (NED) 2 patients had tumour progression 28 (No GH) Mixed 3.4 75% (NED) 2 patients had

tumour progression Buchfelder [98] 55 (GHRT) All NFPA 5 (At least) 20% (NED) 29% had

tumour progression 55 (No GH) All NFPA 5 (At least) 24% (NED) 22% had

tumour progression

Arnold [99] 23 (GHRT) All NFPA 4.6 0% (NED) 35% had

tumour progression 107 (No GH) All NFPA 6.8 0% (NED) 36% had

tumour progression Reference series in patients without GHRT

Gitteos [38] 63 All NFPA 8.1 0% 47%

63 All NFPA 9.1 100% 93%

Brada [94] 252 All NFPA 10.5¤ 100% 94%

Flickinger [101] 112 All NFPA No data 100% 89%

Park [39] 132 All NFPA 3.8¤ 0% 49%

44 All NFPA 5.7¤ 100% 98%

O’Sullivan [95] 126 All NFPA 5.7¤ 0% 49%

GHRT, Growth hormone replacement therapy, NED, No exact data; NFPA, Non-functioning pituitary adenoma;

RT, Radiation therapy; ^, Calculated number; ¤, Follow-up period presented as median value; §, For patients treated with GHRT the mean follow-up time refers to the mean duration of GHRT.

1.7.2 Tumour progression in patients with CP

Craniopharyngiomas have a strong tendency to progress and infiltrate the surrounding

structures, which affects the choice and aggressiveness of tumour treatment. As for

NFPAs, the progression rate for CP is strongly affected by the inclusion of RT in the

primary treatment. The 10-year PFSR in CP patients with initial RT treatment has been

reported to be 83%-92% [27,41,102] (Table 2). In CP patients treated with surgery

alone the 10-year PFSR was 41% for patients with a residual tumour after primary

tumour treatment [27]. The corresponding figure for patients treated surgically without

any residual tumour was 47%-81% [27,36,102].

Table 2. Tumour progression free survival in CP patients sub-grouped depending on GHRT.

Study – First author [ref]

No. of patients

Aetiology Mean follow-up (yrs)

RT Progression free survival rate:

10 years

All patients treated with GHRT §

Clayton [103] 23 All CP 3.8 48% (NED) 7 patients had

tumour progression

Cowell [104] 62 All CP 3.8 No data (NED) 7 patients had

tumour progression

Chung [105] 50 Mixed 3.0¤ 70% (NED) 4 patients had

tumour progression

Darendeliler [106] 1038 All CP 2.8¤ 32% 63%

Comparison between patients with and without GHRT §

Karavitaki [107] 32 (GHRT) All CP 6.3 66% (NED) 4 patients had tumour progression 53 (No GH) All CP 8.3 32% (NED) 22 patients had

tumour progression Rohrer [108] 22 (GHRT) All CP 8.8¤ × 34%× (NED) 11 patients had

tumour progression 7 (No GH) All CP 8.8¤ × 34%× (NED) 4 patients had

tumour progression Müller [109] 54 (GHRT) All CP 2.8¤ 26%× GHRT had no effect on the

progression free survival 60 (No GH) All CP 3.0 26%×

Reference series in patients without GHRT

Van Effenterre [29] * 122 All CP 7.5 6% 60%

Duff [27] * 96 All CP 10× 0% 69%

25 All CP 10× 100% 92%

Stripp [102] * 57 All CP 7.6¤ × 0% 42%

18 All CP 7.6¤ × 100% 84%

Rajan [41] # 173 All CP 12¤ 100% 83%

CP, Craniopharyngioma; GHRT, Growth hormone replacement therapy; NED, No exact data; RT, Radiation therapy;

×, For the entire study; *, No information about GHRT in the paper; ¤, Follow-up period presented as median value;

#, 26 children were treated with GHRT; §, For patients treated with GHRT the mean follow-up time refers to the mean duration of GHRT.

A few studies have been performed to assess whether GHRT influences the risk of tumour progression in CP patients during the past 30 years (Table 2). In one of the first studies, Clayton et al., published their experiences in 1988 after having treated 23 paediatric CP patients with GHRT for a mean follow-up time of 3.8 years and found that seven patients had tumour progression [103]. Almost 20 years later, Darendeliler presented the results of the KIGS (Pfizer International Growth database) registry-study in terms of recurrence of brain tumours in paediatric patients receiving GHRT. In 1038 CP patients with a median duration of GHRT of 2.8 years the 10-year PFSR was 63% [106].

Table 2. Tumour progression free survival in CP patients sub-grouped depending on GHRT.

Study – First author [ref]

No. of patients

Aetiology Mean follow-up (yrs)

RT Progression free survival rate:

10 years

All patients treated with GHRT §

Clayton [103] 23 All CP 3.8 48% (NED) 7 patients had

tumour progression

Cowell [104] 62 All CP 3.8 No data (NED) 7 patients had

tumour progression

Chung [105] 50 Mixed 3.0¤ 70% (NED) 4 patients had

tumour progression

Darendeliler [106] 1038 All CP 2.8¤ 32% 63%

Comparison between patients with and without GHRT §

Karavitaki [107] 32 (GHRT) All CP 6.3 66% (NED) 4 patients had tumour progression 53 (No GH) All CP 8.3 32% (NED) 22 patients had

tumour progression Rohrer [108] 22 (GHRT) All CP 8.8¤ × 34%× (NED) 11 patients had

tumour progression 7 (No GH) All CP 8.8¤ × 34%× (NED) 4 patients had

tumour progression Müller [109] 54 (GHRT) All CP 2.8¤ 26%× GHRT had no effect on the

progression free survival 60 (No GH) All CP 3.0 26%×

Reference series in patients without GHRT

Van Effenterre [29] * 122 All CP 7.5 6% 60%

Duff [27] * 96 All CP 10× 0% 69%

25 All CP 10× 100% 92%

Stripp [102] * 57 All CP 7.6¤ × 0% 42%

18 All CP 7.6¤ × 100% 84%

Rajan [41] # 173 All CP 12¤ 100% 83%

CP, Craniopharyngioma; GHRT, Growth hormone replacement therapy; NED, No exact data; RT, Radiation therapy;

×, For the entire study; *, No information about GHRT in the paper; ¤, Follow-up period presented as median value;

#, 26 children were treated with GHRT; §, For patients treated with GHRT the mean follow-up time refers to the mean duration of GHRT.

A few studies have been performed to assess whether GHRT influences the risk of

tumour progression in CP patients during the past 30 years (Table 2). In one of the first

studies, Clayton et al., published their experiences in 1988 after having treated 23

paediatric CP patients with GHRT for a mean follow-up time of 3.8 years and found

that seven patients had tumour progression [103]. Almost 20 years later, Darendeliler

presented the results of the KIGS (Pfizer International Growth database) registry-study

in terms of recurrence of brain tumours in paediatric patients receiving GHRT. In 1038

CP patients with a median duration of GHRT of 2.8 years the 10-year PFSR was

63% [106].

Three studies have directly compared the progression rate in CP patients with and without GHRT (Table 2). Karavitaki et al. showed a lower frequency of tumour progression in 32 patients with GHRT compared to 53 patients without [107]. The frequency of RT treatment was, however, more than doubled in the GHRT group compared to non-GHRT group. Rohrer and colleagues showed similar frequencies of tumour progression in 22 patients with GHRT compared to seven patients without GHRT [108]. In the most recent paper, Müller and colleagues assembled a multicentre study and recruited 54 childhood-onset CP patients with GHRT and 60 controls. GHRT was not found to be an individual risk factor for tumour progression, but the median duration of GHRT in this study was only 2.8 years [109].

In summary, previous non-comparative studies do not suggest that there is an increase in the tumour progression rate in CP patients receiving GHRT, nor do the few comparative studies of GHRT in CP patients, which have had a limited number of patients or short duration of therapy.

1.8 Mortality and morbidity

A long list of local tumour-related symptoms and symptoms caused by hypopituitarism have been described in patients with non-functioning pituitary tumours, but for many years the long-term effects on mortality was unknown. In 1990, Rosén and Bengtsson showed that patients with hypopituitarism, mainly caused by non-functioning pituitary tumours, had an excess mortality, particularly from vascular diseases [55]. In the years following, two out of three retrospective studies, including between 172 and 348 patients, showed an excess mortality in hypopituitary patients [110-112]. A decade after Rosén and Bengtsson published their paper, Tomlinson and colleagues presented results from the West Midlands Hypopituitary database in which 1014 hypopituitary patients (57% NFPAs, 12% CPs) were studied regarding mortality [49]. This large study also showed an increase in the overall mortality ratio in hypopituitary patients compared to the general population (standardised mortality ratio (SMR) 1.9, 95% confidence interval (CI) 1.6-2.2). Subgrouping by specific aetiologies showed that NFPA and CP had SMRs of 1.7 (95% CI 1.3-2.2) and 9.3 (95% CI 5.8-15), respectively. In NFPA patients, the excess mortality was mainly explained by deaths due to respiratory and vascular diseases. For CP patients, the excess mortality was caused by respiratory and cerebrovascular deaths.

1.8.1 Mortality in patients with NFPA

In Nilsson and colleagues’ registry-based study of 2279 patients with pituitary adenoma, mortality was investigated during the time period 1958 to 1991 [15]. Patients with acromegaly and Cushing’s disease were excluded. Since prolactin measurements were not used in routine clinical care during more than half of the duration of that study, it is likely that a substantial number of prolactinomas was included in the study population. The study reported an excess mortality (SMR 2.0, 95% CI 1.9-2.2) in patients with a pituitary adenoma and a significantly higher mortality in women compared to men. The most common cause of death was cardiovascular disease, including cerebrovascular disease. Later, Lindholm and colleagues were unable to

Three studies have directly compared the progression rate in CP patients with and without GHRT (Table 2). Karavitaki et al. showed a lower frequency of tumour progression in 32 patients with GHRT compared to 53 patients without [107]. The frequency of RT treatment was, however, more than doubled in the GHRT group compared to non-GHRT group. Rohrer and colleagues showed similar frequencies of tumour progression in 22 patients with GHRT compared to seven patients without GHRT [108]. In the most recent paper, Müller and colleagues assembled a multicentre study and recruited 54 childhood-onset CP patients with GHRT and 60 controls. GHRT was not found to be an individual risk factor for tumour progression, but the median duration of GHRT in this study was only 2.8 years [109].

In summary, previous non-comparative studies do not suggest that there is an increase in the tumour progression rate in CP patients receiving GHRT, nor do the few comparative studies of GHRT in CP patients, which have had a limited number of patients or short duration of therapy.

1.8 Mortality and morbidity

A long list of local tumour-related symptoms and symptoms caused by hypopituitarism have been described in patients with non-functioning pituitary tumours, but for many years the long-term effects on mortality was unknown. In 1990, Rosén and Bengtsson showed that patients with hypopituitarism, mainly caused by non-functioning pituitary tumours, had an excess mortality, particularly from vascular diseases [55]. In the years following, two out of three retrospective studies, including between 172 and 348 patients, showed an excess mortality in hypopituitary patients [110-112]. A decade after Rosén and Bengtsson published their paper, Tomlinson and colleagues presented results from the West Midlands Hypopituitary database in which 1014 hypopituitary patients (57% NFPAs, 12% CPs) were studied regarding mortality [49]. This large study also showed an increase in the overall mortality ratio in hypopituitary patients compared to the general population (standardised mortality ratio (SMR) 1.9, 95% confidence interval (CI) 1.6-2.2). Subgrouping by specific aetiologies showed that NFPA and CP had SMRs of 1.7 (95% CI 1.3-2.2) and 9.3 (95% CI 5.8-15), respectively. In NFPA patients, the excess mortality was mainly explained by deaths due to respiratory and vascular diseases. For CP patients, the excess mortality was caused by respiratory and cerebrovascular deaths.

1.8.1 Mortality in patients with NFPA

In Nilsson and colleagues’ registry-based study of 2279 patients with pituitary

adenoma, mortality was investigated during the time period 1958 to 1991 [15]. Patients

with acromegaly and Cushing’s disease were excluded. Since prolactin measurements

were not used in routine clinical care during more than half of the duration of that study,

it is likely that a substantial number of prolactinomas was included in the study

population. The study reported an excess mortality (SMR 2.0, 95% CI 1.9-2.2) in

patients with a pituitary adenoma and a significantly higher mortality in women

compared to men. The most common cause of death was cardiovascular disease,

including cerebrovascular disease. Later, Lindholm and colleagues were unable to

confirm the excess mortality reported by Nilsson et al. in a surgical series of 160 NFPA patients (SMR 1.2, 95% CI 0.9-1.6) [113]. However, after subgrouping regarding to gender, they found an increased mortality ratio in women with NFPA compared to the general population. The mortality in NFPA patients is therefore still somewhat unclear.

In addition, the impact of clinical factors such as surgery, RT, hypopituitarism, DI or age at diagnosis on mortality is largely unknown.

1.8.2 Mortality and morbidity in patients with CP

For patients with CP, Bülow et al. (1998) was the first to report an excess mortality (SMR 5.6, 95% CI 3.7-8.2) in a series of 60 patients, mainly due to cardiovascular causes, including cerebrovascular causes [114]. Since then, two retrospective series of 70 and 54 patients from referral centres have reported an increased SMR of 8.8 (95% CI 5.4-13) and 2.9 (1.4-5.0), respectively [23,115]. In addition, Pereira et al. also reported a possible increase in the incidence of cerebrovascular accidents and myocardial infarctions compared to the general population [115].

Despite the fact that CPs receive a great deal of attention from many physicians, as stated by Dr J. T. Rutka “There is perhaps no other primary brain tumour that evokes more passion, emotion, and, as a result, controversy than does the craniopharyngioma”

[116], there are only three small studies from referral centres that have investigated the mortality in CP patients. As a consequence, the effect of clinical factors such as age at diagnosis, tumour treatment and hormone deficiencies on mortality and morbidity in CP patients are still largely unknown.

confirm the excess mortality reported by Nilsson et al. in a surgical series of 160 NFPA patients (SMR 1.2, 95% CI 0.9-1.6) [113]. However, after subgrouping regarding to gender, they found an increased mortality ratio in women with NFPA compared to the general population. The mortality in NFPA patients is therefore still somewhat unclear.

In addition, the impact of clinical factors such as surgery, RT, hypopituitarism, DI or age at diagnosis on mortality is largely unknown.

1.8.2 Mortality and morbidity in patients with CP

For patients with CP, Bülow et al. (1998) was the first to report an excess mortality (SMR 5.6, 95% CI 3.7-8.2) in a series of 60 patients, mainly due to cardiovascular causes, including cerebrovascular causes [114]. Since then, two retrospective series of 70 and 54 patients from referral centres have reported an increased SMR of 8.8 (95% CI 5.4-13) and 2.9 (1.4-5.0), respectively [23,115]. In addition, Pereira et al. also reported a possible increase in the incidence of cerebrovascular accidents and myocardial infarctions compared to the general population [115].

Despite the fact that CPs receive a great deal of attention from many physicians, as stated by Dr J. T. Rutka “There is perhaps no other primary brain tumour that evokes more passion, emotion, and, as a result, controversy than does the craniopharyngioma”

[116], there are only three small studies from referral centres that have investigated the

mortality in CP patients. As a consequence, the effect of clinical factors such as age at

diagnosis, tumour treatment and hormone deficiencies on mortality and morbidity in CP

patients are still largely unknown.

2 AIM

Overall aim of this thesis was to study tumour progression as well as mortality and morbidity in patients with NFPA and CP.

The specific aims of the thesis were:

Paper I: To study if growth hormone replacement therapy increases the frequency of tumour progression in patients with NFPA.

Paper II: To study if growth hormone replacement therapy increases the frequency of tumour progression in patients with CP.

Paper III: To study mortality in patients with NFPA.

Paper IV: To study mortality and morbidity in patients with CP.

2 AIM

Overall aim of this thesis was to study tumour progression as well as mortality and morbidity in patients with NFPA and CP.

The specific aims of the thesis were:

Paper I: To study if growth hormone replacement therapy increases the frequency of tumour progression in patients with NFPA.

Paper II: To study if growth hormone replacement therapy increases the frequency of tumour progression in patients with CP.

Paper III: To study mortality in patients with NFPA.

Paper IV: To study mortality and morbidity in patients with CP.